Outline

• The need to help scientists improve code reproducibility
• (Environment * Software) + (Code) + (Inputs) = (Outputs)
• Encapsulate = recreate environment + make workspace transparent
• Organize workspace using general best practices
• “Clean” code using minimization and reformatting
• Data provenance = environment details + data processing pipeline

“Software should not limit science.”

Encapsulator

Goal: Simplify computational reproducibility

1. Create a data “capsule” with code, data and environment
2. Increase transparency with “cleaned” code and workspace

Email me: matthewklau@fas.harvard.edu